Method for predicting section curve of heavy ion single event effect of device
Technical Field
The invention belongs to the field of research of a space radiation effect simulation test technology and an anti-radiation reinforcement technology, and relates to a method for predicting a heavy-ion single-event effect cross section curve based on a proton or neutron single-event effect cross section.
Background
Under the space radiation environment, the single event effect is an important factor influencing the reliability of an electronic system of a spacecraft, and heavy ions and protons are main sources causing the single event effect of an electronic device. The proton mainly induces the single particle effect by the ionization and deposition energy of the secondary particle generated by nuclear reaction with the device material in the sensitive area, and the heavy ion induces the single particle effect by the deposition energy of direct ionization in the sensitive area. The ground simulation heavy ion single event effect is generally to utilize a heavy ion irradiation device generated by an accelerator, select more than 5 heavy ion LET value points in an experiment, and obtain a relation curve of a device single event effect section and an LET value, so as to evaluate the single event resistance of the device.
With the improvement of device performance, the increase of integration level and the development of packaging technology, the flip-chip technology has become a main packaging form, and the technology is characterized in that: the device is flipped over on a substrate, connected in the form of solder bumps, with a substrate of several hundred microns thick over the sensitive region of the device. Due to the limited ion energy and range of the heavy ion accelerator, heavy ions are difficult to penetrate through the substrate to reach the sensitive region of the device. Therefore, when the flip device is used for carrying out heavy ion accelerator experiments, the cover of the device is generally required to be opened and thinning treatment is carried out, the device is easily damaged in the operation process, on the other hand, for heavy ions with high LET values and high atomic numbers, the single-nucleus energy of the heavy ions is further reduced, even if the thinned device is difficult to meet the requirements for the ion range in the experiments, and the heavy ion single-particle effect evaluation work of the flip device is greatly difficult. And the middle-high energy proton has low LET value, small energy loss when penetrating through the material and long range, and can effectively penetrate through the device package and the substrate to reach a sensitive region of the device to trigger the single event effect so as to obtain a complete proton single event effect section curve of the device.
Aiming at the problems and the current situation, the method for predicting the heavy-ion single-event effect cross section curve based on the experimental data of the proton or neutron single-event effect is provided by considering the correlation of the proton and heavy-ion single-event effect generation mechanism, the heavy-ion single-event effect cross section curve of the device can be obtained through the experimental data of the proton and the corresponding simulation calculation, and the practical problem that the heavy-ion single-event resistance of the flip-chip device is difficult to evaluate is effectively solved.
The patent application No. 200710177960.5, publication No. CN100538378C entitled "method for obtaining relationship between single particle effect cross section and heavy ion linear energy transfer", provides an experimental method based on heavy ion single particle effect cross section of heavy ion accelerator tester; the patent application No. 202010982765.5, publication No. CN112230081A, entitled "a method for calculating equivalent LET value in pulsed laser single particle effect test" provides a method for using the data of the pulsed laser single particle test to equivalent heavy ions with different LET value single particle effect cross sections. The patent application No. 201711173677.5, publication No. CN108008289B entitled "a method for acquiring proton single event effect cross section" provides a method for acquiring proton single event effect cross section data based on a heavy ion single event effect cross section curve. The three methods do not relate to a method for predicting a heavy ion single event effect cross section curve based on a device proton single event effect cross section.
Disclosure of Invention
The invention provides a method for predicting a heavy ion single event effect cross section curve of a device, which can obtain a complete heavy ion single event effect cross section curve of an inverted device under the condition of not carrying out a heavy ion single event effect experiment, provides an effective technical means for evaluating the heavy ion single event resistance of the inverted device, and overcomes the defect that the heavy ion single event effect evaluation process of the inverted device in the prior art is difficult to realize.
The technical solution of the invention is as follows:
a method for predicting a heavy ion single event effect cross section curve is characterized by comprising the following steps:
the method comprises the following steps: performing proton or neutron single event effect experiment on the device to be tested to obtain a proton or neutron single event effect cross section curve of the device;
step two: carrying out longitudinal cutting analysis on the device, constructing a device structure model, and simulating and calculating a secondary particle LET spectrum generated by nuclear reaction between protons or neutrons with different energies and a device material;
step three: estimating 2 heavy ion single event effect section data points through proton data, and preliminarily fitting a heavy ion single event effect section curve;
step four: respectively carrying out integral calculation on the heavy ion single event effect section curve obtained in the third step and the LET spectrum of the secondary particles under different proton or neutron energies obtained in the second step to obtain proton or neutron single event effect sections under different energies;
step five: and (3) comparing the integral calculation data in the fourth step with the proton or neutron single event effect experiment data in the first step, continuously adjusting the fitting parameters of the heavy ion single event effect section curve when the deviation exceeds a set range, and repeating the fourth step until the deviation between the integral calculation data and the proton or neutron experiment data is within the set range, so that the heavy ion single event effect section curve at the moment is obtained.
Further, the first step specifically comprises:
carrying out proton or neutron single event effect experiment of the device, carrying out Weibull fitting on the obtained proton or neutron experiment data to obtain a fitted proton or neutron single event effect cross section curve sigma
p(E
p):
In the formula, σsat-pIs a saturated cross section of proton or neutron single event effect in cm2;Ep0Is the proton or neutron single event effect energy threshold, in MeV; w is a scale parameter; s is a shape parameter; epIs the proton or neutron energy, in MeV.
Further, the second step is specifically:
2.1) longitudinally cutting and analyzing the device to obtain the thicknesses and material components of the packaging layer, the heat dissipation silicone grease, the substrate and the sensitive layer of the device, and constructing a sensitive volume model of the device;
2.2) calculating the energy as E by using Monte Care particle transport simulation softwarepThe protons or neutrons react with the device material to generate a probability of secondary particles with an LET value L in the sensitive layer, obtaining a probability function p (E)pL) versus LET value.
Further, the third step is specifically:
3.1) preliminarily estimating heavy-ion single-event effect saturation section data points sigma according to the proton single-event effect saturation section combination formula (1-2)sat-ion:
σsat-ion=106×σsat-p (1-2)
3.2) calculating the energy E near the inflection point of the curve of the section of the proton or neutron single event effect in the first Monte Carlo calculation steppAnd counting the probability p (L) that the equivalent LET value is greater than the LET threshold value>L0) And estimating another heavy ion single event effect section data point sigma by the formula (1-3)ion(L):
Wherein L is the equivalent LET value, L0Is the LET threshold;
3.3) combining the two heavy-ion single-particle effect cross sections estimated in 3.1) and 3.2), and preliminarily fitting a heavy-ion single-particle effect cross section curve according to the formula (1-4)
Further, the fourth step is specifically:
the integral expression used in the fourth step is:
in the formula, σp(Ep) Is energy of EpA proton or neutron single event effect cross section of (a); p (E)pL) is energy of EpThe probability that the LET value of the secondary particle is L is generated by the nuclear reaction of the proton or neutron with the device material; sigmaionAnd (L) is a heavy ion single event effect section with an LET value of L.
Further, the device to be tested is a flip chip device.
The invention has the beneficial effects that:
1. the method can obtain the complete heavy-ion single-particle effect cross section curve of the flip device without carrying out a heavy-ion single-particle effect experiment, solves the technical bottleneck that the heavy-ion single-particle resistance of the flip device is difficult to evaluate, and greatly reduces the experiment cost.
2. The proton or neutron single event effect test can be carried out in the air, and the proton or neutron single event effect section test can be carried out without unsealing or thinning the device, so that the damage to the device is reduced, and the test difficulty is lower.
3. The invention starts from the fundamental mechanism that the proton or neutron and the device material generate nuclear reaction to generate secondary particles to trigger single event effect, the physical concept is clear, and the calculation time and the data precision meet the actual application requirements.
Drawings
FIG. 1 is a flow chart of one embodiment of the present invention;
FIG. 2 is a sensitive volume structure model containing device multilayer material information;
FIG. 3 is a plot of probability function of secondary particles generated in a sensitive layer by nuclear reaction of protons of different energies with device materials versus LET value;
FIG. 4 is a preliminarily fitted heavy ion single event effect cross-sectional curve;
FIG. 5 is a comparison graph of proton experimental data versus simulated calculation data for satisfactory deviation;
fig. 6 is a graph comparing the cross-sectional curve of the heavy ion single event effect after final calibration with the experimental data of heavy ions.
Detailed Description
The following embodiments are described in detail with reference to the accompanying drawings, and the present invention is only for illustrative purposes, but not intended to limit the scope of the present invention.
Fig. 1 is a flowchart of a method for predicting a heavy-ion single-event-effect cross-section curve based on a proton single-event-effect cross-section, and the steps of the method are described in detail with reference to fig. 1.
S1, carrying out proton single event effect experiment on the flip FPGA device, carrying out Weibull fitting on experimental data to obtain a fitted proton single event effect cross-sectional function, wherein the expression is as follows:
and S2, performing longitudinal cutting analysis on the flip FPGA, and constructing a sensitive volume structure model of the device according to the longitudinal material process information of the device, which is shown in figure 2. Energy E is calculated by using Mongolian card particle transport simulation software Geant4pThe proton and the device material have nuclear reaction to generate the probability of a secondary particle with an LET value L in the sensitive layer, and a probability function p (E) is obtainedpL) versus LET values, see FIG. 3.
S3 ] the heavy ion single-particle upset saturation section sigma is preliminarily estimated by the formula (1-2)
sat-ion=10
6×σ
sat-p=10
6×2.1×10
-15cm
2/bit=2.1×10
-9cm
2And/bit. The equivalent LET value of 40MeV protons in the sensitive layer is 1.71MeV cm through calculation by adopting Geant4
2Mg, probability p (L) at this time>L
0)=3.17×10
-5(ii) a The LET value of 1.71MeV cm is estimated by combining the formula (1-3) with the proton single event effect section at the position of 40MeV
2The heavy ion single particle upset section of/mg is
. A preliminary heavy ion single event effect cross-sectional curve is obtained from the two points through Weibull fitting, and is shown in FIG. 4, and the expression is as follows:
s4 ] reacting the protons with different energies in the expression (1-7) and S2 in S3 to the device material nucleus to generate the LET spectrum p (E) of the secondary particle in the sensitive layerpAnd L) carrying out integral calculation according to a formula (1-5) to obtain proton single event upset sections under different energies.
And S5, comparing the integral calculation data in the S4 with the S1 proton single event effect experiment data, and repeating the S4 if the deviation is larger, until the deviation between the integral calculation data and the proton experiment data is reduced to a certain range, wherein the heavy ion single event effect cross section curve is the required curve as shown in FIG. 5. Wherein, the heavy ion single event effect cross section Weibull curve expression (1-8) obtained after multiple adjustments:
weibull fitting was performed on the data from the heavy ion single event effect experiment in FIG. 6, and the expression is shown in (1-9):
therefore, the calibration calculation result and the heavy ion single particle experiment result have better consistency.
In another embodiment, a heavy-ion single-event-effect cross section curve based on a neutron single-event-effect cross section predictor can be obtained by the method.